Geographically accurate underground tunnels - Google Maps

[metrailway]

I don't think it has been mentioned here before but Google Maps now, for a couple of months, is showing the geographically correct routes of underground railways in Central London.

However, if you click a tube station, the coloured underground lines shown do not follow this same mapping. I presume this is intentional.

Not very important but might be of interest to some members.

[norbitonflyer]

I'd noticed that too.

In Another Place it was noted that Google and Open Street Map show different alignments for the City branch of the Northern Line between Euston and Camden Town. Google shows it following Hampstead Road and Camden High Street, passing very close to Mornington Crescent station. OSM has it passing about 200m further west .Does anyone know which is correct?

[theblackferret]

Jan 28, 2015 20:49:51 GMT norbitonflyer said:

I'd noticed that too.

In Another Place it was noted that Google and Open Street Map show different alignments for the City branch of the Northern Line between Euston and Camden Town. Google shows it following Hampstead Road and Camden High Street, passing very close to Mornington Crescent station. OSM has it passing about 200m further west .Does anyone know which is correct?

I noticed this, too-useful idea, but can't be sure which is right-Wiki has OSM on it, so it can't help for once.

I'm going to guess OSM, though-informed guess, because the earlier Tube lines had been built to follow under the streets following their line, to avoid costly wayleaves/demolition/compensation. So, as that City Branch was constructed slightly after that, the loopy alignment, which doesn't look on OSM to be following the line of streets at all, would be more feasible.

Wonder if an old Ordnance Survey Map would show them? Have a feeling they used to.

EDIT

Well, from Clive Bilson's site A History of the London Tube Map, you might find this pre-Beck 1924 pocket map useful to decide which one is correct:

1924 map

That looks like it represented what was on (or in this case, under) the ground, rather than the diagrammatic format we all know & love, why else stick a loop like that on the map unless that is how it lays? So, OSM has my vote as correct.

[Deleted]

The OS marked sub-surface (Met & District) tunnels on some maps/scales - take a look here, and explore other scales & dates; but not I think Tube tunnels (try navigating to King William Street.

And take a look at FOI requests here and here

[wimblephil]

I'm sure I read once that the exact routing of the tunnels was not allowed to be publicised due to safety and security risks, and that actually there were no documents in existence that even showed the accurate positing of them.

Or is that all just belloni... !?

[rincew1nd]

Well there was that Old St incident where the developers drilled the supports for their new building into the tunnel; they were criticisied for not looking at any maps!

[longhedge]

Jan 29, 2015 6:08:07 GMT wimblephil said:

I'm sure I read once that the exact routing of the tunnels was not allowed to be publicised due to safety and security risks, and that actually there were no documents in existence that even showed the accurate positing of them.

Or is that all just belloni... !?

I have an old Barts Atlas of London, and it appears to show the actual tunnel routing. For example, at Shepherds Bush (Central) it clearly shows the strange Caxton Curve routing. 

Unless I am missing something, I fail to grasp what is the security risk in showing the routing. The tunnels are many feet underground, and I am not sure how a person(s) with sinister intentions would realistically be able to do anything on the surface to compromise  the integrity of the tunnel deep underground. 

[rheostar]

Jan 29, 2015 6:08:07 GMT wimblephil said:

I'm sure I read once that the exact routing of the tunnels was not allowed to be publicised due to safety and security risks, and that actually there were no documents in existence that even showed the accurate positing of them.

Or is that all just belloni... !?

You're right, the publication of the routing is restricted, but the rest is all just belloni! ;-)

[John Tuthill]

Jan 29, 2015 8:23:04 GMT rincew1nd said:

Well there was that Old St incident where the developers drilled the supports for their new building into the tunnel; they were criticisied for not looking at any maps!

And don't forget the incident at Redbridge(?) where the piling bit actually hit a train!

[theblackferret]

Jan 29, 2015 9:40:40 GMT longhedge said:

Jan 29, 2015 6:08:07 GMT wimblephil said:

I'm sure I read once that the exact routing of the tunnels was not allowed to be publicised due to safety and security risks, and that actually there were no documents in existence that even showed the accurate positing of them.

Or is that all just belloni... !?

I have an old Barts Atlas of London, and it appears to show the actual tunnel routing. For example, at Shepherds Bush (Central) it clearly shows the strange Caxton Curve routing. 

Unless I am missing something, I fail to grasp what is the security risk in showing the routing. The tunnels are many feet underground, and I am not sure how a person(s) with sinister intentions would realistically be able to do anything on the surface to compromise  the integrity of the tunnel deep underground. 

I suspect any security considerations here were during and after WWII. There was one rather pathetic scare in WWI re the abandoned tunnels at King William Street being used by German spies, but the police soon gave the nutters behind that the bird for their stupidity.

If you think about it, there was always a far greater chance of a German invasion second time around, not to mention more accurate bombing with high explosives. Two very good reasons to keep information to a minimum, and why that 1924 map I found is likely to be unaffected by that sort of thing, and therefore accurate.

As to nowadays, well, I don't envy TfL's security staff at all.

[Chris M]

In either the late 90s or early 2000s the central London bus map showed the routings of the Underground tunnels.

[norbitonflyer]

I can't see that knowing where deep tube tunnels run is much of an issue - they're not exactly easy to get into from above - at least not on purpose. The subsurface ones, which are much more accessible, as seen here are shown quite accurately

Jan 29, 2015 10:31:51 GMT John Tuthill said:

And don't forget the incident at Redbridge(?) where the piling bit actually hit a train!

Actually the train hit the drill bit. Fortunately it was only a surveyor's drill taking core samples, and not a piling bit like the one at Old Street. The driver was lucky it wasn't a couple of feet further to the left though.

Jan 29, 2015 11:39:21 GMT theblackferret said:

[ German ....... [WW2] more accurate bombing with high explosives. .

WW2 bombing was far from accurate. Hitting a target that narrow by simply dropping something from several hundred feet in the dark whilst dodging ack ack fire was simply not possible. If you blindfold someone and then get them to throw a hundred darts in the general direction of a dartboard, some of them will hit it. There is even a reasonable chance of hitting a double.

The direct hits on Balham and Bank stations, for example, were not the result of precision targetting but were, simply, just two of the thousands of sites in London hit by bombs in WW2. Although above-ground railways were indeed targetted, (and occasionally hit, although lineside buildings were just as likely to cop it - this church being such a case) the chances of a tube station being hit were no greater than any other location in London.

Jan 29, 2015 12:01:42 GMT Chris M said:

In either the late 90s or early 2000s the central London bus map showed the routings of the Underground tunnels.

I don't think these were particularly accurate, if only because the tunnels so often run directly underneath the bus routes themselves!

[phillw48]

My 5 year old London A-Z shows the sub surface tunnels as dotted lines.

[North End]

Jan 28, 2015 20:49:51 GMT norbitonflyer said:

I'd noticed that too.

In Another Place it was noted that Google and Open Street Map show different alignments for the City branch of the Northern Line between Euston and Camden Town. Google shows it following Hampstead Road and Camden High Street, passing very close to Mornington Crescent station. OSM has it passing about 200m further west .Does anyone know which is correct?

I haven't been able to load either of the maps quoted, however I can say from civils plans I've seen that the city branch passes much closer to Mornington Crescent than 200m.

On the city branch there are two empty construction tunnels, one for each tunnel, which head off east towards Mornington Crescent station. From the plans it appears there were working shafts which actually went up between the two Mornington Crescent platforms, the city branch being slightly deeper at this point. The shafts are now backfilled, which presumably happened as soon as construction work finished. The empty connecting tunnels I would estimate are no more than about 25 metres. Today there is no connexion between the branches at Mornington Crescent, and as far as I can tell there never was, as the shafts would have gone straight to street level.

There was also a similar working shaft which surfaced at Ampthill Square, off Hampstead Road. This was also backfilled, although evidence of the site's location can be seen at track level. Similar features can be seen at various locations on the Northern Line if one knows where to look.

[norbitonflyer]

Jan 29, 2015 23:57:16 GMT North End said:

I haven't been able to load either of the maps quoted, however I can say from civils plans I've seen that the city branch passes much closer to Mornington Crescent than 200m.

here's a linkie to the Google version

here's OSM
www.openstreetmap.org/export#map=17/51.53480/-0.14055

The locations of the working shafts suggests OSM has got it wrong

[wimblephil]

Jan 29, 2015 9:40:40 GMT longhedge said:

Jan 29, 2015 6:08:07 GMT wimblephil said:

I'm sure I read once that the exact routing of the tunnels was not allowed to be publicised due to safety and security risks, and that actually there were no documents in existence that even showed the accurate positing of them.

Or is that all just belloni... !?

I have an old Barts Atlas of London, and it appears to show the actual tunnel routing. For example, at Shepherds Bush (Central) it clearly shows the strange Caxton Curve routing. 

Unless I am missing something, I fail to grasp what is the security risk in showing the routing. The tunnels are many feet underground, and I am not sure how a person(s) with sinister intentions would realistically be able to do anything on the surface to compromise  the integrity of the tunnel deep underground. 

I dunno, it's just something I read once! I wish I could remember where now, I may have mis-read but that's what I took from it!

[phillw48]

Jan 30, 2015 0:16:04 GMT norbitonflyer said:

Jan 29, 2015 23:57:16 GMT North End said:

I haven't been able to load either of the maps quoted, however I can say from civils plans I've seen that the city branch passes much closer to Mornington Crescent than 200m.

here's a linkie to the Google version

here's OSM
www.openstreetmap.org/export#map=17/51.53480/-0.14055

The locations of the working shafts suggests OSM has got it wrong

The first map takes me to my local area not part of London and there is no means of moving the map around. This has happened quite often just recently so perhaps Google have changed something?

[trt]

Try this link instead:

goo.gl/5266te


BBCode was breaking the link at the comma, so I used Google's URL shortener to eliminate them.

[rincew1nd]

I've fixed the links by telling the forum software what a link was, rather than letting it figure it out itself.

[longhedge]

In the last week I was looking from the web at a detailed large scale plan of Tottenham Court Road station, showing passageways, where the Central Line/Northern Line tunnels run, plus the very long Crossrail tunnels in the south with the actual street plus pavements etc overlaid. so you can ascertain exactly where everything is when on the surface. So there does not appear to be an embargo if this is published.

[phillw48]

I have a map produced in 1944 showing all tunnels and subways beneath the City of London including the Post Office Railway. This is one of a series of maps produced by the City planners for post war development most of which was never carried out. Fortunately in the case of a proposed dual carriageway ring road around the City.

[murph9000]

As far as the precise location and security goes, it's really extremely pointless to try to hide that information. In well under 60s, I can think of a fairly easy technique to accurately map the tunnel positions in 3D with a briefcase of easily obtainable electronics and simply sitting on the trains as they run through the tunnels in normal service. Out of respect for the forum, I won't go into specifics on just how to do it, only say that some of the technology required is commonly available in high end smart phones and tablets. Some moderately clever data processing might be required, and it would probably take a few journeys to collect high precision data, but it's really not difficult in principle. Additionally, "security through obscurity" has long been recognised as dangerously useless.

[crusty54]

The Crossrail website features the alignment of the tunnels and gives a good representation of the alignments with other lines.

[longhedge]

I have the complete suite of Video 125 Underground `Drivers Eyes views`. which covers most of the Underground, albeit only in one direction. I would have felt if there was a real security risk, filming would never have been authorised.

Perhaps related, I have just downloaded a 125 page presentation of the London Bridge Station rebuild from the web, which has very detailed plans, detailing exactly where the room from which the Security Cameras are monitored, plus power supplies etc.

[theblackferret]

Herewith another map, from 2004. This is the physical features of London with actual physical Tube alignments superimposed on it:


mappy

That would indicate the Google alignment around Mornington Crescent is wrong.

[Deleted]

Feb 8, 2015 18:55:38 GMT theblackferret said:

Herewith another map, from 2004. This is the physical features of London with actual physical Tube alignments superimposed on it:


mappy

That would indicate the Google alignment around Mornington Crescent is wrong.

Looking at how well aligned (not) Tube alignments and railways are when both are are on the surface, I don't think the tube (underground) alignments are at all accurate.

[Deleted]

Looking at the TfL web site Transparency page there is here a link to the Property Asset Register, which shows (if you zoom in - warning, it can be a bit slow), inter alia, the 'LUL Zone of Influence' where the tube lines run, against OS large (1:1250) scale mapping.

[Antje]

One of the reasons why OSM is struggling to have the tube alignments 100% correct is the copyright clearance. OpenStreetMap's deliberate challenge is that all the data has to come from original (survey, photo, GPS) or compatible sources (UK Open Government License).

Maybe you could be the guys to help with getting the OSM tube alignments right, but TfL has to clearly provide their alignment data under the UK Open Government License and not the Ordnance Survey OpenData License (see wiki.openstreetmap.org/wiki/Licensing/Ordnance_Survey_OpenData_License).

Even when it comes to successfully getting the alignments under the UK OGL, we still have to follow the import guidelines at wiki.openstreetmap.org/wiki/Import/Guidelines. We also have to publish evidence of UK OGL clearance to quell future issues from other parties. Such procedures may take some time depending on how busy the regular mappers are.

[Deleted]

Here is the vector-based test-harness I use for viewing the OpenStreetMap tube data:

• www.paul.sladen.org/osm/lul.html

Click the (+) in the top right to select the "Mapnik" map underlay.

Yes, as @Antje sums it up: it's not just a question of knowing "where", but of locating "where" from legally usable sources. This means either survey + GPS (which doesn't work in tunnels), or textual descriptions, or out-of-copyright materials, or in the longer-term negotiated open data releases under a suitable licence. There are lots of points where the OpenStreetmap tunnel alignment is known to need tweaking and many of these are listed on:

• wiki.openstreetmap.org/wiki/London_Underground#Known_errata

What the errata are waiting for is known-good legally-usable information to make the updates.

The Northern Line CX routeing was not "random" per-se but was drawn carefully from the following textual description received from an (otherwise) reliable source [with whom most of you are familiar with]:

"Sorry about the delay in replying, but the relevant material was archived away ... From Camden Town, the Bank branch runs across the street pattern. It runs under the intersection of Delancey Street and Arlington Road, then the intersection of Albert Street with Mornington Street, Mornington Terrace and Mornington Place, just skims Stanhope Street where it meets Granby Terrace, cross Harrington Street just south of the intersection with Mackworth Street, cross Hampstead Road just north of Varndell Street, then Cardington Street twice, once immediately east of Hampstead Road and then in St.James' Gardens, where Robert Street would meet it if it extended that far east."

In the absence of anything usable to contradict it, the lines based on this description have been left in-situ and had not been adjusted to a better alignment—it may have been known to be wrong, but somebody needs to publicly say that so that it can be cited as the source.  I have now (23 February 2015) cited the statement on this thread as the source and re-routed the Northern Line Bank Branch up West Hampstead Road; this follows other situations—

• www.openstreetmap.org/way/230065328#map=16/51.5302/-0.1361

Fixing these little bits is where yourself and others could help greatly—one simply needs to suggest textually (ie. without reference to any copyrighted maps!) what precisely needs to be moved where, and they can be fixed quite quickly!

@etr220: Part of goal for independently surveying Tube/shaft locations is not only to get /better/ data, but to show that the information is already widely known in the public knowledge from first principles and daily usage. For example there is another FOI request from somebody else in 2011 that was refused on the basis of the secrecy:

• www.whatdotheyknow.com/request/location_of_mid_tunnel_vent_shaf

By demonstrating that maps already exist, it should dispel the argument that the information is "secret", and then it mainly becomes an issue of delivery format and of licensing. On the licensing front, the Ordnance Survey have last week laid-down their own OS Open Data Licence and switched to standard Open Government Licence, which helps and may help to bring successful negotiation forward.

@rincew1nd: In the Moorgate/Northern City drilling incident the tunnel route was shown on OpenStreetMap, but not on Ordanance Survey or other maps. TfL stated that whilst TfL didn't have a tunnel in the area (the tunnel had been transferred...) but that the conveyancing solicitor would need to contact Network Rail (which didn't happen).

@longhedge: Yes, I have used the RAIB report on the Old Street incident as example of where making usable data widely available actively contributes to public safety, rather than endangering it. The first recommendation from the Rail Accident Investigation Branch (on page 36) is that:

"Railway Infrastructure Managers … should take all reasonable steps to publicise this information"

This may/may not have contributed to the TfL "Public Access Property Asset Register" finally being put online. I shall try not to taint myself too much, but I did spot that the location of PiccEx Junction (north-west) of Heathrow Central is shown outside the "LUL Zone of Influence"—and once you've spotted one error in a map it stops being completely trustworthy anyway. …Which is why OpenStreetMap is mapped "from scratch"!

@murph9000: A better Inertial Measurement Unit (IMU/IMS) [such as found on aircraft] would be an excellent contribution to mapping all sorts of things where GPS isn't reliable.

• Sladen, Paul (July 2013) "Tube to Chunnel: Underground mapping" (video, 20 minutes)
  

I did this "Tube to Chunnel" talk a couple of years ago, much of which covers an overview of London Underground mapping in respect of OpenStreetMap, in which I noted out-of-copyright deposited plans weren't always accurate and encouraging validation with IMU/gyro systems. By example, the corner Wood Lane to Uxbridge Road is currently mapped on OSM per the parliamentary deposited plans, however AFAIK the limits of deviation were used to their maximum extent after negotiation and a more gentle curve taken after the present-day location of the control centre.

@phillw48: If the 1944 tunnel map of London is Crown Copyright (term of 50 years), this map would now be out-of-copyright and so it would be excellent to scan and contribute as source material; particular the Post Office railway which could do with some significant refining.  Please do scan and upload somewhere, including any markings in the margins to validate its publication date and origin.

For the sub-surface The c.1897 OS 1:1,056 NLS scans should provide a clean source for refining the sub-surface lines (particularly Mark Lane/Tower Hill/Crescent and the area under Barbican) eg. maps.nls.uk/geo/explore/#zoom=18&lat=51.50996&lon=-0.07753&layers=163

@crusty54: Crossrail are another organisation where it's been hard in communicating to convey the difference between an item being in the public knowledge and being legally usable (out-of-copyright/public domain/explicit reuse allowed. I had put in FOI request to try and obtain the GIS data explicitly under the Open Government Licence for OSM use—it is of course already available from a public Crossrail WMS server, but under unclear terms:

• www.whatdotheyknow.com/request/open_data_crossrail_route_shapef

(We can try again when the tunnels have broken-through as, by having relied on Section 22 (future publication), Crossrail have also committed to publication at some point).

In the middle of the timeline are the Victoria/Jubilee/JLE which are too recent for out-of-copyright map sources and so OGL-licensed open data centre-lines from TfL would be a good solution.

@theblackferret: The you've found appears to be just Bézier curves joining dots, see for example the Bakerloo and Northern line being the wrong way around between Embankment and Lambeth North.

So to return to @norbitonflyer's original starting point: The deposited plans are not geographically precise; TfL's published plans are not totally precise; the current Ordnance Survey plans are incomplete; Metro Carto isn't perfect (I've sent in corrections); the Google Maps tunnel routes are nowhere near accurate; pretty much every non-OSM "geographical map" I've seen is wrong: and can be eliminated with quick checks for South Ken Bends, Embankment?North Lambeth, Claxton Curve, E&C overrun, Paddington Bakerloo and Baker Street Jubilee arrangements.

…Yes, the OpenStreetMap centre-lines are not perfect either (yet), but they can be slowly refined and improved, and are probably the most accurate "legally reusable" ones available.  Plus, improving the station and tunnel mapping is something that we (all, collectively) could actively get involved with, either by mapping or by going out with a pen-and-paper, or collecting numbers/distances only (such as VP numbers—eg. the shafts at Henley's Roundabout and Whitfield Place so they can be added), or merely textually describing known errors and actual route corrections based on first-hand knowledge.

Having the map at the end to point to during discussion will be immensely rewarding, and useful in these conversations!

[Deleted]

One issue over the 'under the Barbican' area is that a stretch was realigned as part of the Barbican Centre construction, c1964.

Another potential issue to beware of is that, while the 1897 OS maps are out of copyright, the NLS (or anyone else's) scans may themselves not be (as a 'new work' when the scanning was done). NB: I have not checked the NLS's licence rubric, nor am I a lawyer.